Use of ecotoxicity test and ecoscores to improve the management of polluted soils: case of a secondary lead smelter plant

HAL Id: hal-00815325

https://hal.archives-ouvertes.fr/hal-00815325

Submitted on 18 Apr 2013

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Use of ecotoxicity test and ecoscores to improve the management of polluted soils: case of a secondary lead

smelter plant

Yann Foucault, Marie-José Durand, Karine Tack, Eva Schreck, Florence Geret, Thibaut Leveque, Philippe Pradère, Sylvaine Goix, Camille Dumat

To cite this version:

Yann Foucault, Marie-José Durand, Karine Tack, Eva Schreck, Florence Geret, et al.. Use of ecotoxicity test and ecoscores to improve the management of polluted soils: case of a secondary lead smelter plant. Journal of Hazardous Materials, Elsevier, 2013, Vol. 246-247, pp. 291-299.

�10.1016/j.jhazmat.2012.12.042�. �hal-00815325�

This is an author-deposited version published in: http://oatao.univ-toulouse.fr/

Eprints ID: 8490

To link to this article: DOI: 10.1016/j.jhazmat.2012.12.042 URL: http://dx.doi.org/10.1016/j.jhazmat.2012.12.042

To cite this version: Foucault, Yann and Durand, Marie-José and Tack, Karine and Schreck, Eva and Geret, Florence and Leveque, Thibaut and Pradère, Philippe and Goix, Sylvaine and Dumat, Camille Use of ecotoxicity test and ecoscores to improve the management of polluted soils: case of a secondary lead smelter plant. (2013) Journal of Hazardous Materials, Vol. 246-247 . pp. 291-299. ISSN 0304-3894

Open Archive Toulouse Archive Ouverte (OATAO)

OATAO is an open access repository that collects the work of Toulouse researchers and makes it freely available over the web where possible.

Any correspondence concerning this service should be sent to the repository

administrator: staff-oatao@listes-diff.inp-toulouse.fr

Use of ecotoxicity test and ecoscores to improve the management of polluted soils: case of a secondary lead smelter plant

Yann Foucault

, Marie-José Durand

, Karine Tack

, Eva Schreck

, Florence Geret

, Thibaut Leveque

, Philippe Pradere

, Sylvaine Goix

, Camille Dumat

aUniversitédeToulouse,INP-ENSAT,Avenuedel’Agrobiopôle,31326Castanet-Tolosan,France

bUMR5245CNRS-INP-UPS,EcoLab(Laboratoired’EcologieFonctionnelle),Avenuedel’Agrobiopôle,BP32607,31326Castanet-Tolosan,France

cSTCM,SociétédeTraitementsChimiquesdesMétaux,30AvenueFondeyre,31200Toulouse,France

dUniversitédeNantes,UMR6144CNRSGEPEA,DépartementGénieBiologique,18BoulevardG.Defferre,85000LaRochesurYon,France

eINERIS,ParcTechnologiqueAlata,BP2,60550Verneuil-en-Halatte,France

fUMR5563CNRS/UPS/IRD/CNESGET,14AvenueEdouardBelin,31400Toulouse,France

gUMRCNRS5602,LaboratoireGEODE,PRESUniversitédeToulouse,CentreUniversitaireJean-Franc¸oisChampollion,PlacedeVerdun,81012AlbiCedex9,France

g r a p h i c a l a b s t r a c t

Keywords:

Sustainablemanagementofpollutedsoils Metaltraceelements

Ecotoxicity Landfilling Leaching

a b s t r a c t

Withtheriseofsustainabledevelopment,rehabilitationofbrownfieldsiteslocatedinurbanareashas becomeamajorconcern.Managementofcontaminatedsoilsinrelationwithenvironmentalandsanitary riskconcernsisthereforeastrongaimneedingthedevelopmentofbothusefultoolsforriskassessment andsustainableremediationtechniques.Forsoilspollutedbymetalsandmetalloids(MTE),thecriteriafor landfillingarecurrentlynotbasedonecotoxicologicaltestsbutontotalMTEconcentrationsandleaching tests.Inthisstudy,theecotoxicityofleachatesfromMTEpollutedsoilssampledfromanindustrialsite recyclinglead-acidbatterieswereevaluatedbyusingbothmodifiedEscherichiacolistrainswithlumines- cencemodulatedbymetalsandnormalizedDaphniamagnaandAlivibriofischeribioassays.Theresults wereclearlyrelatedtothetypeofmicroorganisms(crustacean,differentstrainsofbacteria)whosesensi- tivityvaried.Ecotoxicitywasalsodifferentaccordingtosamplelocationonthesite,totalconcentrations andphysico-chemicalpropertiesofeachsoil.Forcomparison,standardleachingtestswerealsoper- formed.PotentiallyphytoavailablefractionofMTEinsoilsandphysico-chemicalmeasureswerefinally performedinordertohighlightthemechanisms.Theresultsdemonstratedthattheuseofapanelof microorganismsissuitableforhazardclassificationofpollutedsoils.Inaddition,calculatedeco-scores permittorankthepollutedsoilsaccordingtotheirpotentiallyofdangerousness.Influenceofsoiland MTEcharacteristicsonMTEmobilityandecotoxicitywasalsohighlighted.

∗Correspondingauthorat:Ecolab(LaboratoireEcologieFonctionnelleetEnvironnement),INP-ENSAT,Avenuedel’Agrobiopôle,BP32607,AuzevilleTolosane,31326 Castanet-Tolosan,France.Tel.:+33534323903;fax:+33534323901.

E-mailaddress:camille.dumat@ensat.fr(C.Dumat).

1. Introduction

Originallylocatedontheoutskirtsofcities,numerousindustrial sites,sometimesabandoned,arenowinurbanareasandarethere- forelikelytohaveenvironmentalandhealthriskstosurrounding populations[1,2].Currently,rehabilitationofthesitesfrequently entailsexcavationofpollutedsoils[3].Excavatedsoilscanthus followtwodifferentways:landfilling,expensiveandenergyinten- sive,orreuse/recycling,integratedtosustainabledevelopment.The choiceofaspecifictrackmainlydependsontotalandleachable concentrationsofthepollutantinthesoil[2].Amongthenumer- ouspollutantsobservedinurbanandperi-urbanareas,tracemetals areoftenpresentinsoils[4];atmosphereemissionsbysmelters beingoneof themain anthropogenicsource [5,6].MTEspecia- tionandcompartmentalizationinsoilscanmodifytheirimpacton livingorganisms[5].Now,numerouspublicationsconcludedthat thesetwoparametersarestronglyinfluencedbysoilorganicmatter (OM)content,pHandtexture[7–9].AccordingtoMatejczyketal.

[10],chemical weatheringofsoilmineralsfavoursMTEsolubili- zationandleachatesproduction.Then,theseleachatescanpollute surroundingsoilsandwaters.Accordingtothecouncil directive n^◦1999/31/CE,leachingtestswithchemicalanalysisaretherefore currentlyusedfortheassessmentofenvironmentalhazardsofpol- lutedsoils.But,landfillingisofteninevitableforstronglypolluted soils,withhigh“hazardlevel”(assessedbyleachedandtotalMTE concentrations).Moreover,accordingtoFoucaultetal.[1],profes- sionalsconsiderthethresholdsetastoorestrictiveandtheyregret thatexcavatedsoilsarealmostalwaysmanagedaswaste.

InadditiontothemeasureoftotalandleachedMTEconcentra- tions,itappearsthereforethatknowledgeofMTEavailability[11]

andecotoxicitymaycarryusefulinformation[12–14]toimprove environmentalriskassessment[10].Actually,theaccurateestima- tionofmetalphytoavailabilityinpollutedsoilsandsolidwastes, usingsinglechemicalextraction[15]carryinterestingdatatoper- formpertinentriskassessmentandremediationefforts[16,17].Soil qualityintegratesbothphysicochemicalandbiologicalcharacter- istics[18].Moreover,accordingtoPlazaetal.[14]microorganisms playimportantrolesin numeroussoilfunctions.Soilsareoften pollutedwitha largevarietyof compoundsleading topossible interactions[19],thusasreviewedbyKimandOwens[20]studyof leachatesecotoxicityprovidesadirectfunctionalcharacterization ofvariouspollutantmixtures.But,onlyfewstudiesconcerntheuse ofecotoxicologicalteststomonitorcontaminationandbioremedi- ationefficiencyofpollutedsoils[21]andnewtestsarerequired byindustrialsitesmanagerstoassessenvironmentalrisks.Among them,microbialbioassays offerquick,cheapand easyecotoxic- ity(toxicityandmutagenicity)andbioavailabilitymeasurements onbacteria[22,23].However,inmanycases,microbialbioassays cannotbedirectlyusedfortheidentificationandquantificationof compoundsduetothelackofspecificityoftheengineeredmicroor- ganisms [24] and further studiesare neededto improve these biotests.

Theaimofthisstudywasthereforetoassesstheecotoxicity ofleachatesforlandfillingofMTEcontaminatedsoilsbyvarious complementary biotests, in addition to usual physicochemical measures.Moreprecisely,thefollowingtwoscientificobjectives wereaimed:(1)whatisthepertinenceofecotoxicityteststoassess amorerealistichumanexpositiontocontaminatedsoilleachates?

(2) What is the influence of soil physicochemical parameters onMTEmobility and leachatesecotoxicity?Severalstudiesuse specificbacteria straintosense thepresence of metals insoils [25–28], nevertheless, the development of statistical model to understandthelinkbetweenchemicalconcentrationofcompound andbacteriasensorsisstillon-goingwork[23].So,theoriginality ofthisstudywastocombinetheuseofnewbacterialstrainsnever testedinacontextoftheremediationofanindustrialpollutedsite

and calculationof eco-scores which facilitatesthecomparisons betweendifferentsoils.

2. Materialsandmethods

2.1. Soilsamplingandpreparation

AccordingtoWongetal.[29],themostrelevantsoillayerto studytheenvironmentaland sanitaryimpacts ofMTEin urban areas is between 0 and 25cm. Ten top soil samples (Fig. 1) werethereforecollectedinthecourtyardoftheChemicalMetal TreatmentsSociety(STCM),asecondaryleadsmelterwhichcur- rently recycles batteries locatedin theurban areaof Toulouse (43^◦38^′12^′′N,01^◦25^′34^′′E). Thisplantwas chosenbecause ofits activityandurbanlocation,andmanydataarealreadyavailable [1,4–6,30].Thesedataalloweddefiningdifferentareasintermsof environmentalandsanitaryrisksthatcanvaryaccordingtopast andpresentactivities.Moreover,previousstudiesoftheparticles releasedin theatmosphereby Uzuetal. [5]and Schrecketal.

[4]revealedthepresenceofseveralMTE(Pb,As,Cu,Cd,Znand Sb)andgaveinformationonthemainleadspeciation:PbS,PbSO4, PbO PbSO4,a-PbOandPb(byorderofabundance).Allsoilsamp- lingpointsarepresentedinFig.1;theyweredried,sievedunder 2mmandtreatedintriplicate.

2.2. Physico-chemicalanalysis

pH,organicmatterand limestonecontents, cation exchange capacity(CECMetson)andtexture,weredeterminedforallsoil samples,respectively,accordingtothenormsISO10390[31],ISO 10694[32],ISO10693[33],NFX31-310[34]andISO11277[35].

Pb,As,Cu,Cd,ZnandSbtotalconcentrationsweremeasuredby ICP-OES(IRISIntrepidIIXXDL)aftermineralizationinaquaregia accordingtoISO11466[36](HNO365%,HCl37%,ratio1:3(v/v)).

ThedetectionlimitsofPb,Cd,Sb,As,CuandZnwere0.3,0.2,0.2, 0.2,1.3and2.2mgL⁻¹,respectively,whereasthelimitsofquan- tificationwereabout0.4,0.3,0.4,0.3,2and3mgL⁻¹,respectively.

Theaccuracyofmeasurementswascheckedusingacertifiedrefer- encematerial141R(BCR,Brussels).Theconcentrationsfoundwere within95–102%ofthecertifiedvaluesforallmeasuredelements.

2.3. Leachingtests

Normalizedleachingtest[37]wasappliedtoallsoilsamples.

This procedureconsisted of a single extraction with deionised water,usingasolid-to-liquidratioof1/10.10gofsoil(granulome- tryatleast<4mmaccordingtothenorm)wasmixedwith100mL deionisedwaterduring24hwithend-over-endagitationat5rpm.

Aftercentrifugationat3000×gduring15min,theleachateswere filteredwithcellulose0.45mm(Millipore^®)filters.10mLofeach leachateswerethenacidifiedwithHNO365%priortoanalysisby ICP-OES(IRISIntrepidIIXXDL,analyticalerrors<5%).Theotherpart ofleachateswasnotacidifiedsoasnottodisturbmicroorganisms usedforfurtherecotoxicologicaltests.

2.4. MTEphytoavailabilityestimate

PotentiallyphytoavailableMTEconcentrationswereestimated byCaCl2extractionsaccordingtoUzuetal.[5].In25mLpolypro- pylenecentrifugationtubes,10mLof10⁻²MCaCl2wereaddedto 1.0gofsoil.Theliquidtosolidratioof10ishighenoughtoavoid samplesheterogeneities[38].Afteragitationend-over-endduring 2hat5rpmat20^◦C,sampleswerethencentrifugedduring30min at10,000×g.Supernatantwassievedthrougha0.22mmmeshand acidifiedat2%withHNO3 (15N,suprapur99.9%).MTEconcen- trationswerefinallymeasuredbyICP-OES(IRISIntrepidIIXXDL,

Fig.1.Situationoftheindustrialstudysite,locationandcharacteristicsofthe10samplingpoints.

analyticalerrors<5%).Ahousereferencesoilwasusedtoquality controlofCaCl2extraction:thissoildescribedbySchrecketal.[4]

hastheadvantageofpresentingthesametypeofcontamination thatthesoilsstudiedinthiswork.Actually,itisasoilhistorically pollutedbybatteryrecyclingemissions([Pb]=1650±20mg;kg⁻¹).

Using thatreference soil,thedetection limitsof Pb,Cd,Sb, As, CuandZnwere0.3,0.2,0.2,0.2,1.3and2.2mgL⁻¹,respectively, whereasthelimitsofquantificationwereabout0.4,0.3,0.4,0.3,2 and3mgL⁻¹,respectively.Theconcentrationsfoundwerewithin 95–102%ofthereferencevaluesforallmeasuredelements.

2.5. Ecotoxicityassessment 2.5.1. Daphniamagna

Afirstacutetoxicityofleachateswasperformedonthewater fleaD.magna(Origin,lessthan24hold)accordingto[39].Four replicatesweretestedforeachsoilsolutionandfiveneonateswere usedineachreplicate,with10mLoftestsolution.Organismswere fed2hbeforebutnotduringtheexperiment.Aparafilmstripwas thenputonthemultiwallplateplacedintheincubatorat20^◦C indarkness.ThemobilityofD.magnawasrecordedafter24hand 48h,andinhibitionratewascalculated.

2.5.2. Microtox^®

Firstusedtoassessacuteecotoxicityofmetalsinaquaticmedia [40]andnormalizedsince2007[41],solid-phaseMicrotox^®testis nowcurrentlyusedtoevaluatethetoxicityofcontaminatedsoils orsediments[4,42,43].TheMicrotoxtestmeasuresthedecrease inlightemittedbythebioluminescentbacteriaVibriofischeriafter 5,15and30minofexposure[43].Inviewofevaluatingtoxicity ofcollectedsoils,theMicrotox81.9%BasicTestwiththeinstru- mentMICROTOXM500purchasedfromR-Biopharm(France)was used.100mLofrevitalizedbacteria(Lot10J1010A)wereaddedto eachMicrotox^®tube,gentlymixedwithapipette.900mLofeach leachatewastransferredintotheglasscuvetteintheMicrotox^® analyzerandallowedtoequilibratefor5minbeforereading[44].

Lightemissionwasrecordedandtheoutputdataanalyzedusing Microtox^® Omni software Version 1.18[45]. All samples were testedintriplicate.Bacteriavalidityandtheset-upofthemeasure- mentprocedureswereverifiedbyreferencetoxin(ZnSO4,7H2O) accordingtotheISO11348regulationspecifications[41].

2.5.3. Inductionofbioluminescentbacteria

Many bacterial sensors are dedicated to the specific detec- tion of pollutantsor pollutant family.These used bio-elements arebacterialstrains(Escherichiacoli)geneticallymodified.In all cases,reportergenesluxCDABEarecloneddownstreamofapro- moterallowingtohighlightthespecificorsemi-specificpresence forcertaincompoundsinasample[46,47].Inthecase ofmetal detection,thepromotersusedaremostlyinvolvedinthemech- anismsof bacterialresistancetoheavymetals[48–50].A setof fivebioluminescentbacterianamelyE.coliTaclux,E.coliZntlux, E.coliArslux,E.coliCopluxand E.coliMerlux wasusedinthis study(Table1).Bacterialgrowthandlyophilizationwererealized accordingtoJouanneauetal.[24].Atthebeginningofthebioas- say,thelyophilizedbacteriawerereconstitutedwith100mLper wellofdistilledwaterfor30minat+30^◦C.25mLofleachatewas addedtoeachwell,andafterwardthemicroplatewasincubated for60minat+30^◦C.Monitoringofbioluminescencewasrecorded using a microplate luminometer (Microlumat Plus Lb96V). The resultswereexpressedbythelogarithmoftheinductionratioorthe inhibitionratefortheinduciblestrainsandtheconstitutivestrain, respectively.Theinductionratio(IR)wascalculatedasfollows:

- IR=(RLU·s⁻¹)i-IR/(RLU·s⁻¹)0-IR. (RLU·s⁻¹)i-IR is the biolumines- cence afterinduction with a sample, and (RLU·s⁻¹)0-IR is the backgroundluminescence.Theinhibition rate(InR) wascalcu- latedasfollows:

-InR=1−(RLU·s⁻¹)i-InR/(RLU·s⁻¹)0-InR.(RLU·s⁻¹)i-InRisthebiolu- minescenceafterexposurewithasample,and(RLU·s⁻¹)0-InRis thebackgroundluminescence.

Table1

ObservedMTEanddetectionlimitsforthedifferentE.colistrains[24].

MTE Detectionlimitofbacterialstrains(standarddeviation)(mM)

Zntlux Arslux Merlux Coplux

Cd 0.0045(0.0003) 5.9(2.3) 0.011(0.002) –

Hg 0.01(0.005) – 1.2×10⁻⁷(1×10⁻⁷) –

As^(III) 28.52(7.1) 0.256(0.0014) 15.6(4.3) –

Cu 16.92(2.9) – – 90.5(11.7)

Pb 2.2(0.6) 4.16(0.8) – –

Sn 12.95(4.24) – – –

As^(V) 9.32(1.24) 0.3(0.06) 12.65(5.4)

Zn 1.7(0.62) – 2.3(0.14) –

Ni 4.4(1.6) – – –

Co 0.22(0.014) – – –

Cr^(VI) 597.2(121.3) – – –

Ag – – – 2.75(0.11)

Fe 4.34(0.48) – 16.1(7.6) –

Mn – – – –

Decisiontreesweredesignedfromthelearningsetofbacterialbio- luminescencedatausingthesoftware“Metalsoft”.Theyarespecific toonlyonecompoundandorganizedinseveralbinarybranches.

Eachbranchsplitsdataaccordingtothevaluesofonlyonevari- able:inductionrationorinhibitionrateobtainedfromonestrain.

Theprocesscontinuesuntilthetargetvalueisobtained[24].

2.6. Statisticalanalysis

Alltestswereperformedintriplicateandresultsarepresented asmean±SD(standarddeviation).Thestatisticalsignificance of valueswascheckedusingananalysisofvariance(ANOVA)usingthe Statistica9.0packagesoftware.EachMTE-extractedconcentration (bothbywaterandCaCl2procedures)wascomparedtorespective totalconcentration.Significantdifferences (p-value<0.05) were measuredbytheLSDFishertest.

3. Results

3.1. Physico-chemicalcharacteristics

Soil propertiesarereportedin Table2.Thesephysicochemi- calcharacteristicssignificantlydifferinfunctionofsampleorigin:

itmeanslocalizationontheindustrial siteinrelationwithpro- cess. pH value varied between 6.9 and 9.2, CEC value varies between2.6 and10.5cmol(+)kg⁻¹ and amounts of soilorganic matter and carbonates (CaCO3) were highly variable, respec- tively, from 0.9 to 46.7gkg⁻¹ and from 0 to 15.0gkg⁻¹. MTE concentrationsinpollutedsoilsamples(Table3)werealsovery heterogeneous:maximumleadconcentrationis42,400mgPbkg⁻¹ and other elements are also present at high levels (up to 2095mgSbkg⁻¹, 288mgAskg⁻¹, 286mgCukg⁻¹, 294mgZnkg⁻¹ and80.9mgCdkg⁻¹).Alltheseconcentrationswereclearlyabove thenationalgeochemicalbackgroundasshowninTable3.

3.2. Chemicalextractions

MTE(Pb,As,Cu,Cd,ZnandSb)concentrationsweremeasured forallthesolutionsobtainedbyperformingthethreeextractions (aquaregia,waterandCaCl2)onthestudiedpollutedsoilsamples (Tables3–5;Figs.2and 3).LeachedMTEamountsinwater and correspondingratios(incomparison withaquaregiaextraction, consideredas “total”)weresignificantly depending onelement nature(Table 4a and Fig. 2). The highest extracted concentra- tionswererecordedforlead,antimonyandzinc(MTEwithhigh totalconcentrations),respectively,152.5,158and9mgkg⁻¹,i.e.

8.7%,7.3%and7.9%.Incomparison,copper(atequivalentcontent) wassignificantlylessextractedthanzinc.Althoughquantitatively

lowextracted(≤5.4mgCdkg⁻¹),Cdwasproportionallyoneofthe mostwater-solubleelement(upto15.9%forS6).Arsenicwasthe less extracted MTE witha maximum concentration reachedof 2.2mgAskg⁻¹(7.5%ofthetotalforS4).

CaCl2extractionsresults(Table4bandFig.3)showedseveral contrastedbehavioursdependingbothonchemicalelementand soilproperties.ThehighestleadquantitiesextractedbyCaCl2were observedforS1,S3andS7(upto178.5mgkg⁻¹).However,forS4

samplewithhightotalleadconcentration,theextractedfractionis low(1.3mgkg⁻¹).Conversely,antimonyextractedconcentration reached306.6mgSbkg⁻¹.OtherMTEshoweda low extractabil- ity(intermsofquantityandratio)whateverthesample,except

Fig.2.Ratios(%)betweenMTEleachedaccordingtotheEN12457-2procedure andaquaregiaextraction,forthe10soilsamples:(a):sumoftheratiosofleached- concentrations(in%)<10%and(b)sumoftheratiosofleached-concentrations(in

%)>30%.

Table2

Mainphysicochemicalcharacteristicsofthetensoilsamples.

Sample pHwater Organicmatter(gkg⁻¹) CEC(cmol(+)kg⁻¹) LimestoneCaCO3(gkg⁻¹) Granulometry(%)

Clay Thinsilt Roughsilt Thinsand Roughsand

S1 7.0 31.4 8.9 8.0 13.3 18.3 14.2 15.7 38.5

S2 8.7 9.6 3.7 7.0 3.6 10.7 10.9 19.7 55.1

S3 6.7 14.0 6.0 0.0 9.2 13.7 12.6 19.7 44.8

S4 8.7 12.3 8.9 15.0 7.8 12.1 11.4 21.3 47.3

S5 9.2 0.9 2.6 7.0 2.7 4.3 4.8 2.9 85.3

S6 9.0 1.6 3.3 4.0 4.3 8.1 6.7 15.5 65.4

S7 6.9 46.7 10.5 0.0 6.7 9.4 8.7 21.6 53.5

S8 7.5 3.4 3.3 4.0 3.3 5.9 4.1 8.6 78.0

S9 8.5 6.0 6.9 4.0 10.4 17.3 11.2 19.5 41.6

S10 8.9 1.3 3.3 8.0 1.5 5.6 4.4 7.3 81.2

Table3

AquaregiaMTEconcentrationsforthe10soilsamples(mineralizationaccordingto[29]).

MTEconcentrations(mgkg⁻¹) NGB^a S1 S2 S3 S4 S5 S6 S7 S8 S9 S10

Pb 9–50 39,800 1425 42,400 37,250 1020 297 35,700 1445 1750 1065

As 1–25 288 28.7 51.5 52.5 5.8 9.3 34.3 8.65 13.3 10.2

Cu 2–20 286 14.7 143.5 116 13.6 16.1 249 19.4 59.5 60.5

Cd 0.05–0.45 18.4 2.24 34.3 4.15 3.39 0.69 80.9 3.39 4.7 11.3

Zn 10–100 294 37.1 216 218 42.8 41.9 545 55 116 94

Sb 0.2–10 2095 53.5 1555 2175 23.5 13.1 1955 15.9 44.5 9.15

aNGB:NaturalGeochemicalBackgroundinFrance.

Table4

Leachedandphyto-availableMTEconcentrations(mgkg⁻¹)inthetenpollutedsoilsamples.

Leachedconcentrations(mgkg⁻¹) S1 S2 S3 S4 S5 S6 S7 S8 S9 S10

(a)LeachedTEconcentrations(mgkg⁻¹)

Pb 127.8 63.0 126.2 85.6 53.4 51.8 51.7 11.1 152.5 14.9

As 0.29 2.17 0.28 0.33 nd^a 0.38 nd nd 0.54 nd

Cu 1.50 0.71 1.20 0.57 1.44 1.67 0.52 nd 5.04 0.87

Cd 0.52 0.21 1.06 nd 0.21 0.11 5.75 0.05 0.39 0.10

Zn 3.71 1.90 4.19 1.26 3.32 3.86 9.04 0.41 9.13 1.00

Sb 10.0 3.69 9.63 158.2 0.81 2.30 1.64 0.23 3.13 0.22

CaCl2-extractedconcentrations(mgkg⁻¹) S1 S2 S3 S4 S5 S6 S7 S8 S9 S10

(b)Phyto-availableMTEconcentrationsassessedbytheCaCl2procedure

Pb 178.5 0.71 162.9 1.30 0.52 0.60 89.4 14.7 0.31 3.64

As nd nd nd 0.17 nd nd nd nd 0.04 0.01

Cu 0.43 0.20 0.37 0.38 0.19 0.18 0.32 0.21 0.29 0.64

Cd 3.16 0.20 4.16 0.22 0.31 0.22 20.1 0.68 0.22 0.24

Zn 2.77 nd 3.51 0.06 0.20 nd 25.6 0.44 0.08 0.10

Sb 1.54 0.48 1.56 306.6 0.16 0.28 0.77 0.08 0.35 0.11

and:notdetected.

forS7whichregisteredpronouncedpoolsofCdandZnassociated withhightotalconcentrations.Fig.3aandbshowsthefractionof theextractedelementinrelationtototalconcentration.Cadmium appearedasthemostpotentiallyphytoavailableelement.SbandZn alsorepresentedhighextractedfractions,respectively,forS4and S7.Moreover,comparedtotheaquaregiafraction,theCaCl2fraction remainedlower,exceptforthemostpotentiallyphytoavailableCd element(2–32%).

3.3. Ecotoxicitytests

EcotoxicityofleachatesmeasuredbytheinhibitionofD.magna mobilitywashighlyvariable(Table5).Whateverthesampletested, theinhibitionofdaphniamobilityincreasedbetween24hand48h, exceptforS3whoseinhibitionwasnear100%afteronly24h.Eco- toxicitywasalsomaximal(i.e.100%)forS1,S2,S7andS8after48h;

whilethelowerinhibitionwasobservedforS5(15%).Ecotoxicity

Table5

ResultsoftheDaphniamagnaandMicrotox^®testsforthe10leachates.

Ecotoxicitytest Sample

S1 S2 S3 S4 S5 S6 S7 S8 S9 S10

D.magna^a 24h 90 35 100 40 10 25 80 65 45 5

48h 100 100 100 65 15 80 100 100 65 70

Microtox^®b 5min 29 nt^c nt nt nt nt 63 nt nt nt

15min 78.5 nt nt 12 nt nt 92 nt nt nt

30min 93 nt 34 34 nt nt 96 nt nt nt

aInhibitionofmobility(%).

b Inhibitionofbioluminescence(%).

c nt:nottoxic.

Fig.3.Ratios(%)betweenMTEextractedbyCaCl2andaquaregia,forthe10soil samples.(a)Pb,Sb,As,ZnandCuand(b)Cd.

wasnotsimplydependantofMTEconcentration:(i)themostMTE- enrichedleachateswerenotalwaysthemoretoxicant;(ii)leachate ofS8hadlowMTEconcentrationswhiletheinhibitionofdaphnia mobilitywas100%.

ThemeanEC50-30minvalueobtainedforzincsulphatehep- tahydrate(expressedasZn²⁺)was2.38mgL⁻¹,allowingconcluding that the invertebrates lot fulfilled the validation specifications according to [34]. Microtox^® test results (Table 5) showed an increaseinthenumberoftoxic sampleswiththecontacttime:

inhibitionofbioluminescencewasdetectedintwosamplesatthe beginningoftheexperimentandforfourofthemattheend(S1and S7to5min;S1,S4andS7to15min;S1,S7,S4andS3at30min).The measuredecotoxicityalsoincreasedovertimeandwasabove90%

forS1andS7after30minofcontact.Thebacteriaweremostaffected byS7withaninhibitionoftheluminescenceof63%(5min).Unlike thetestonDaphnia,S2,S6,S8,S9andS10showednotoxicity,asS5in bothbioassays.Asdescribedabove,S1,S3,S4andS7areamongthe mostcontaminatedleachates([Pb]>80mgkg⁻¹;[Sb]>10mgkg⁻¹ (exceptS7))(Table4a).

Thesensitivityandspecificityofinduciblebacteriaweremea- suredafter60minincontactwithleachates.Noneofthesample inducedtheluminescenceofCopluxstrain.Onlytwosamples,S1

andS7,showedaslighttoxicityasdemonstratedbytheinhibition ofluminescenceoftheconstitutivestrainpBtaclux(Table6a).

ThesesamplesalsoinducedtheluminescenceofZntlux,Arslux andMerluxstrains.ForS1andS7themaximumIRwasrecorded for Arslux (IR=99.6) and Merlux (IR=138.1), respectively. S3

increasedmoderatelytheluminescenceofMerlux(IR=12.9)and Zntlux (IR=4.0) while S2 induced only Arslux (IR=324.9). The analysis with decision trees was then used to determine the elementspotentiallyresponsibleforecotoxicityofS1,S2,S3 and S7.Crossesbetweenresultssuggestedthepresenceofarsenicin thesefourleachates (up to10⁻⁵M,i.e. more than by chemical

analysis),cadmiumforS1,S3 andS7,biologicallyatlevelslower than thosemeasured chemically(Table6b).Analysis ofS7 also showedthepresenceofcopperandmercury(from10⁻⁴to10⁻⁵M and10⁻⁹to10⁻⁵M).Accordingtotheprevioustests,theseresults alsoconcludedtotheecotoxicity ofS1 and S7,and, toaminor extent,theecotoxicityofS3andS4.

4. Discussion

4.1. MobilityandphytoavailabilityofMTE

Inthisstudy,soilpHwerebasicorclosetoneutralconditions andleachingprocedureslightlyreducedthepHbywateraddition.

Conversely,CaCl2isalreadyknowntonotmodifysoilpHandgive resultscloserfromfieldreality[15].Thus,hazardproposedclassi- ficationofpollutedsoilsdiffersbetweenwaterleachingandCaCl2

procedures.SeveralstudiesalreadyshowedthatMTEextractabil- ityis stronglyinfluenced bythenatureoftheextractingagent, whichcancontrolelementmobility[15,16].Moreover,according toDumatetal.[51]orFerrarietal.[52],solid-liquidMTEtrans- fersduringchemicalextractionsarecomplexreactionsinvolving numerousfactorsthatcaninfluenceMTEspeciationandrelease.

Contacttimeschosenforchemicalextractionswere24hforwater andonly2hforCaCl2inaccordancewiththecommonlyusedpro- tocols:thesetwoprocedurescarrycomplementaryinformationbut theresultsarenotdirectlycomparable.

AllMTEwereextractedinsubstantiallyequalproportionswith water(from0to18%);MTEconcentrationsinCaCl2extractsand correspondingratios,variedintherangeofthosereportedinthe literature[5,58,61],andCdwasthemostavailableelement(upto 32%).Extractedconcentrationswerestrongly correlatedtototal concentrations: R²=0.92 and 0.95 (p<0.0001), respectively, for waterand CaCl2.Atthereverseside,for alltheotherelements norelevant correlationwas foundbetweentotaland extracted fractions.In agreementwithpreviouspublications [5,15],these resultshighlightedtheinfluenceofsoilpropertiesandMTEnature onitsbehaviour.DifferencesobservedinfunctionofMTEnature can be explained by different OMor CaCO3 soil contents, CEC orsoilpH.Insoils,cadmiumisgenerallyeasytodissolvewhich explainitsrelativelyhighextractability[53].Highcorrelationfac- tors wereobserved between exchangeableCu and Znfractions and soil organicmatter amount: R²=0.82 for Cu (p<0.05) and R²=0.91forZn(p<000.1).Theseelementswerethuslessmobile becauseoftheiraffinityforthissoilfraction[11,54].Concerning leadbehaviour,norelationshipwasfoundbetweenextractedand totalconcentrations,andtheinfluenceofevenonesoilparame- terwasdifficulttohighlight.Nevertheless,low extractionratios comparedtothemostconcentratedsamples(S1,S3,S4andS7)can beexplainedbystrongerboundsonsoilphasesasmineralfrac- tion[29].Finally,sorptionofmetalloids asAsand Sb,is mainly controlledbymineralphases[55].ThehighSbamountextracted fromS4couldbeexplainednotonlybyahighertotalconcentration butalsobythehighestCaCO3content[56].Sbcouldbesolubilized undertheinfluenceofsoilbio-physico-chemicalparameterscon- trollingitssorption[57–59].pHandCECwerealreadydescribedas influentparametersofelementextractability[60],retentionand mobilityinsoils[11].Thus,accordingtotheoriginofsoilsamp- ling,differencesinsoilparameterswereobserved(Fig.1):inthe industrialsite,areasnotcoveredorinfiltrationzoneswerethemost impactedbyMTE.TheirorganicmattercontentandCECwerealso thehigher,thusconfirmingtheirroleinsorption/desorptionmech- anisms.Thechoiceoftheextractantistherebyanimportantstep toberelevantinriskassessmentandtoavoidanunder-orover- estimationofphytotoxicity.Finally,thedataobtainedbychemical testsaredifficulttointerpretbecauseofthemanyparametersinter- act.Therealizationofecotoxicityteststomeasuretheimpactof

Table6

Ecotoxicityresultsofbioluminescenceemittedbythebacterialstrains.

Sample Bacterialstrain

ZntLux^a Arslux^a Merlux^a pBtaclux^b

(a)Inductionfactorandinhibitionratecalculatedfromthebioluminescence

S1 4.8 99.6 9.9 6.5

S2 0.9 324.9 0.8 –

S3 4.0 1.2 12.9 –

S4 1.4 0.8 0.9 –

S5 0.7 0.9 1.0 –

S6 0.6 0.8 0.9 –

S7 5.4 79.8 138.1 18.4

S8 1.2 0.9 0.9 –

S9 1.3 2.1 0.8 –

S10 1.2 1.2 0.8 –

Sample Chemicalanalysis Biologicalanalysis(predictionwithdecisiontrees)

Cd As Cu Hg Cd As Cu Hg

(b)ComparisonbetweenMTEleachedwater-solubleconcentrationsandrange“biologically”detectedbythebacterialstrains(unit:M)

S1 4.5×10⁻⁷ 3.7×10⁻⁷ 2.3×10⁻⁶ – 10⁻⁸to10⁻⁷ 10⁻⁶to10⁻⁵ – –

S2 1.8×10⁻⁷ 29.7×10⁻⁷ 1.1×10⁻⁶ – – 10⁻⁶to10⁻⁵ – –

S3 9.1×10⁻⁷ 3.6×10⁻⁷ 1.8×10⁻⁶ – 10⁻⁸to10⁻⁷ <10⁻⁶ – –

S7 48.9×10⁻⁷ 3.3×10⁻⁷ 0.8×10⁻⁶ – 10⁻⁸to10⁻⁷ 10⁻⁶to10⁻⁵ 10⁻⁴to10⁻⁵ 10⁻⁹to10⁻⁵

aIF:inductionfactor.

b Inhibitionrate(%).

pollutiononecosystemsseemsthereforeparticularlyappropriate inthistypeofstudy.

4.2. Relevanceofecotoxicitytestsforriskassessmentposedby landfilling

TheD.magnaecotoxicitytestwasmoresensitivetoMTEimpact thantheMicrotox^®test.But,unliketestsonthedifferentbacterial strains,theydo not bothprovide informationonMTEquantifi- cation.Ecotoxicitydifferencesweremeasuredforsomesamples, especiallyS2andS8.Thesedifferencescanbefirstlyexplainedby waterfleasensitivity.Detectioncapabilitiesoftheecotoxicityof theleachateareactuallydependentonthetestused[62,63]andit hasbeenalreadyshownthatV.fischeriwasgenerallylesssensitive thanD.magna[10,64].Insteadofthesetests,experimentsbyusing

bacterialstrainsallowedtodetermineandquantifytheelement whichwaspotentiallybioavailableand/ortoxicforbacteria[50].

Then,responseinecotoxicitytestswasnotalwaysdirectlycorre- latedwithtotalorwater-solubleconcentrations[65].Theseresults areinagreementwithdatapreviouslyobtainedbyPlazaetal.[14]

concerningtheinfluenceofpHandCEConMTEbehaviourinsoils.

Resultsofthisstudyhaveshownthatthisnewbioassayenables thescreeningofsamplesin termsofenvironmentalrisk during remediationprocess[24].However,thedrawbackofthelackof specificityofonestrainandtheeffectofamixtureofMTE(syner- gisticorantagonisticeffects)couldbeovercomebyusingapanelof bacterialstrainscoupledwithapredictivemodel[24].Duetothe lackofspecificbacteriaforlead,theintroductionofotherstrains induced bylead likeRastonia Metallidurans AE 1433[66] could improvetheinterpretationofthedata.

Table7

Hazardclassification,adaptedfrom[10,60].

Sample Score Class^a MCW^b W^c PW^d

D.magna Microtox Zntlux Arslux Merlux pBtaclux

S1 4 3 1 3 2 2 V 4 2.5 62.5

S2 4 0 0 4 0 0 V 4 1.3 33.3

S3 4 1 1 1 2 0 V 4 1.5 37.5

S4 2 1 1 0 0 0 III 2 0.7 33.3

S5 0 0 0 0 1 0 II 1 0.2 16.7

S6 3 0 0 0 0 0 IV 3 0.5 16.7

S7 4 3 2 3 4 2 V 4 3.0 75.0

S8 4 0 1 0 0 0 V 4 0.8 20.8

S9 2 0 1 1 0 0 III 2 0.7 33.3

S10 2 0 1 1 0 0 III 2 0.7 33.3

a

Class MCW

I Noacutetoxicity PE<20% IR/InR<1 0 II Slightacutetoxicity 20%≤PE<50% 1≤IR/InR<5 1 III Acutetoxicity 50%≤PE<75% 5≤IR/InR<50 2 IV Highacutetoxicity 75%≤PE<100% 50≤IR/InR<100 3 V Veryhighacutetoxicity PE≥100% IR/InR≥100 4

b MCW:maximumclassweightscore.

c W:classweightscore.

d PW:classweightscoreinpercent.

!

!

!

!

1.1. Hazard classification according to ecotoxicity

!

According to Persoone et al. [67] and Matejczyk et al. [10], the samples were ranked into one of five classes on the basis of the percentage effect (PE) found in Daphnia and Microtox^®tests. Rank- ing was based on induction/inhibition rates for bacterial strains. A weight score was calculated for each hazard class to indicate the quantitative importance (weight) of the ecotoxicity in that class.

The weight score was expressed as percentage.

!

),

all test scores

Acknowledgment

!

ANRT (National Agency for Research and Technology) and STCM (Metal Chemical Treatments Society) are acknowledged for their financial support and technical help.

!

References

!

[2] P.A.W. Van Hees, K. Elgh-Dalgren, M. Engwall, T. Von Kronhelm, Re-cycling

= Class weight score =

number of tests performed (

!

(1)

6) of remediated soil in Sweden: an environmental advantage? Resour. Conserv.

Recy. 52 (12) (2008) 1349–1361.

[3] S. Colombano, A. Saada, V. Guerin, P. Bataillard, G. Bellenfant, S. Beranger, D.

Hube, C. Blanc, C. Zornig, I. Girardeau, Quelles techniques pour quels traite-

Class weight score (%) = maximum class weight score × 100 (2)

!

That classification system aimed at the integration of ecotoxic- ity data obtained in a battery of bioassays as describe by Lors et al. [68]. The classification system is based on two values:

a ranking in five acute toxicity classes and a weight score for each toxicity class. The classification of the samples tested in the investigation is reported in Table 7. Samples were classi- fied as slightly and highly toxic in 10%, toxic in 30%, and very highly toxic in 50%. The percentage of class weight class was above 5% for only S7and S1(75% and 62.5%, respectively). These samples were definitely considered as the most hazardous and acutely toxic to the microfauna. The final classification of ecotox- icity risks was S7> S1> S3> S2= S4= S9= S10 > S8> S5= S6. Although the toxicity of some samples (S2 and S8, for instance) could be different depending on the test used, the ranking based on total concentrations and leachable contents of MTE was almost the same; samples S1, S3, S4 and S7 presenting the greatest risks while the less contaminated areas were generally the less hazardous. However, as our results demonstrated that only a small fraction of total MTE soil concentrations can be solubi- lized and phytoavailable; ecotoxicity measures complete therefore efficiently standard performed tests for a realistic risks assess- ment of MTE-contaminated soils accordingly to [10,14]. Moreover, the use of eco-scores improves the comparisons between bioas- says and suggests the use of a restricted battery of tests to perform a cost-effective risk assessment of MTE-contaminated soils.

!

!

2. Conclusion

!

Biotests and eco-scores improve standard tests performed to assess risk on ecosystems induced by polluted soils. In par- ticular, modified bacteria strains sensitive to metals are useful tools highlighting the presence of different MTE and the influ- ence of soil parameters that lead to synergistic or antagonistic effects. Moreover, eco-scores calculation allows an easy and cheap screening of a large number of polluted soil samples and suggests a restricted battery of bioassays to perform a cost-effective risk assessment.

Nevertheless, further researches are required to develop a panel of bacterial strains specific of each MTE. Moreover, the chemi- cal characterization of leachates could provide additional relevant insight with the use of modelling software for determination of MTE speciation and thus to evaluate the predominant parameter influencing soil availability and ecotoxicity. Finally, the procedure developed in the present work could be used in the context of bioremediation techniques like phytoremediation performed on polluted soils, to assessment their efficiency and favour their use in a context of sustainable development.

ments – analyse coûts-bénéfices, BRGM-RP-58609-FR, 2010.

[4] E. Schreck, Y. Foucault, F. Geret, P. Pradère, C. Dumat, Influence of soil ageing on bioavailability and ecotoxicity of lead carried by process waste metallic ultrafine particles, Chemosphere 85 (10) (2011) 1555–1562.

[5] G. Uzu, S. Sobanska, P. Pradère, C. Dumat, Study of lead phytoavailability for atmospheric industrial micronic and sub-micronic particles in relation with lead speciation, Environ. Pollut. 157 (4) (2009) 1178–1185.

[6] X.S. Luo, S. Yu, X.D. Li, Distribution, availability, and sources of trace metals in different particle size fractions of urban soils in Hong Kong: implications for assessing the risk to human health, Environ. Pollut. 159 (2011) 1317–1326.

[7] C. Fritsch, P. Giraudoux, M. Coeurdassier, F. Douay, F. Raoul, C. Pruvot, C.

Waterlot, A. de Vaufleury, R. Scheifler, Spatial distribution of metals in smelter- impacted soils of woody habitats: influence of landscape and soil properties, and risk for wildlife, Chemosphere 81 (2010) 141–155.

[8] F. Vega, M. Andrade, E. Covelo, Influence of soil properties on the sorption and retention of cadmium, copper and lead, separately and together, by 20 soil hori- zons: comparison of linear regression and tree regression analyses, J. Hazard.

Mater. 174 (1–3) (2010) 522–533.

[9] M. Sammut, Y. Noack, J. Rose, J. Hazemann, O. Proux, M. Depoux Ziebel, E. Fiani, Speciation of Cd and Pb in dust emitted from sinter plant, Chemosphere 78 (4) (2010) 445–450.

[10] M. Matejczyk, G.A. Plaza, G. Nale˛ cz-Jawecki, K. Ulfig, A. Markowska-Szczupak, Estimation of the environmental risk posed by landfills using chemical, micro- biological and ecotoxicological testing of leachates, Chemosphere 82 (2011) 1017–1023.

[11] M.A. Kashem, B.R. Singh, T. Kondo, S.M. Imamul Huq, S. Kawai, Comparison of extractability of Cd, Cu, Pb and Zn with sequential extraction in contaminated and non-contaminated soils, Int. J. Environ. Sci. Technol. 4 (2) (2007) 169–176.

[12] CEN 14735, Characterization of Waste. Preparation of Waste Samples for Eco- toxicity Tests, European Committee for Standardization, Brussels, 2006, 47 pp.

[13] I. Buj, J. Torras, M. Rovira, J. de Pablo, Leaching behaviour of magnesium phos- phate cements containing high quantities of heavy metals, J. Hazard. Mater.

175 (1–3) (2010) 789–794.

[14] G.A. Plaza, G. Nale˛ cz-Jawecki, O. Pinyakong, P. Illmer, R. Margesin, Ecotoxic- ological and microbiological characterization of soils from heavy-metal and hydrocarbon-contaminated sites, Environ. Monit. Assess. 163 (2010) 477–488.

[15] E. Meers, R. Samson, F.M.G. Tack, A. Ruttens, M. Vandegehuchte, J. Van- gronsveld, M.G. Verloo, Phytoavailability assessment of heavy metals in soils by single extractions and accumulation by Phaseolus vulgaris, Environ. Exp.

Bot. 60 (3) (2007) 385–396.

[16] N.W. Menzies, M.J. Donn, P.M. Kopittke, Evaluation of extractants for estima- tion of the phytoavailable trace metals in soils, Environ. Pollut. 145 (1) (2007) 121–130.

[17] M. Arshad, J. Silvestre, E. Pinelli, J. Kallerhoff, M. Kaemmerer, A. Tarigo, M.

Shahid, M. Guiresse, P. Pradère, C. Dumat, A field study of lead phytoextrac- tion by various scented Pelargonium cultivars, Chemosphere 71 (11) (2008) 2187–2192.

[18] J. Römbke, A.M. Breure, C. Mulder, M. Rutgers, Legislation and ecological quality assessment of soil: implementation of ecological indication systems in Europe, Ecotoxicol. Environ. Saf. 62 (2005) 201–210.

[19] S.A. Mansour, M.F. Gad, Risk assessment of pesticides and heavy metals con- taminants in vegetables: a novel bioassay method using Daphnia magna Straus, Food Chem. Toxicol. 48 (1) (2010) 377–389.

[20] K.R. Kim, G. Owens, Potential for enhanced phytoremediation of landfills using biosolids – a review, J. Environ. Manage. 91 (2010) 791–797.

[21] L. Leitgib, J. Kalman, K.K. Gruiz, Comparison of bioassays by testing whole soil and their water extract from contaminated sites, Chemosphere 66 (3) (2007) 428–434.

[22] M.J. Durand, G. Thouand, T. Dancheva-Ivanova, P. Vachon, M. DuBow, Specific detection of organotin compounds with a recombinant luminescent bacteria, Chemosphere 52 (1) (2003) 103–111.

[23] T. Elad, E. Benovich, S. Magrisso, S. Belkin, Toxicant identification by a luminescent bacterial bioreporter panel: application of pattern classification algorithms, Environ. Sci. Technol. 42 (22) (2008) 8486–8491.

[24] S. Jouanneau, M.J. Durand, P. Courcoux, T. Blusseau, G. Thouand, Improvement of the identification of four heavy metals in environmental samples by using predictive decision tree models coupled with a set of five bioluminescent Bacte- ria, Environ. Sci. Technol. 45 (7) (2011) 2925–2931.